Estimation of Smart Home Thermophysical Parameters Using Dynamic Series of Temperature and Energy Data

Abstract

A development of the modern methods dedicated to the energy optimization systems for a smart home is being widely studied by community of academic researches and commercial R\&D facilities. To contribute a new method to the existing approaches designed for the understanding of smart home energy losses caused by environmental conditions, an algorithm for estimation of the thermal parameters of the smart home is presented in this paper. Proposed method is based on solving an inverse thermophysical problem in variational formulation, where the temperature and energy data are used as input parameters. This type of problem has been obtained by a reduction of the differential form of a dynamic heat transfer equation to discrete representation in order to use sensor time series data. The realization and testing of developed method has been carried out by using an open-access dataset provided by REFIT SMART Home project. These smart home sensor data have been undergone a cleansing with the subsequent matching with discrete form of the heat transfer problem. Finally, as a result of this process, the effective thermophysical parameters - heat capacity and conductivity have been calculated with the use of solving the inverse problem by one of the state-of-art minimization algorithms.